FASTHash: FPGA-based High Throughput Parallel Hash Table
AI/Machine Learning/Deep Learning
TimeTuesday, June 23rd9:00am - 9:30am
LocationAnalog 1, 2
DescriptionHash table is a fundamental data structure that provides efficient data store and access. It is a key component in AI applications which rely on building a model of the environment using observations and performing lookups on the model for newer observations. In this work, we develop FASTHash, a “truly” high throughput parallel hash table implementation on FPGA on-chip memory. Contrary to state-of-the-art hash table implementations on CPU, GPU, and FPGA, the parallelism in our design is data independent, allowing us to support p parallel queries (p > 1) per clock cycle via p processing engines (PEs) in the worst case. Our novel data organization and query flow techniques allow full utilization of abundant low latency on-chip memory and enable conflict free concurrent insertions. Our hash table ensures eventual consistency - inserts from a PE are visible to all PEs with some latency. We provide theoretical worst case bound on the number of erroneous queries (true negative search, duplicate inserts) due to eventual consistency. We customize our design to implement both static and dynamic hash tables on state-of-the-art FPGA devices. Our implementations are scalable to 16 PEs and support throughput as high as 5360 million operations per second with PEs running at 335 MHz for static hashing and 4480 million operations per second with PEs running at 280 MHz for dynamic hashing.